Download Persistence of invading gypsy moth populations in the United States

Oecologia (2006) 147: 230–237
DOI 10.1007/s00442-005-0271-5
POPULATION ECOLOGY
Stefanie L. Whitmire Æ Patrick C. Tobin
Persistence of invading gypsy moth populations in the United States
Received: 28 January 2005 / Accepted: 8 September 2005 / Published online: 10 December 2005
U.S. Government 2005
Abstract Exotic invasive species are a mounting threat to
native biodiversity, and their effects are gaining more
public attention as each new species is detected. Equally
important are the dynamics of exotic invasives that are
previously well established. While the literature reports
many examples of the ability of a newly arrived exotic
invader to persist prior to detection and population
growth, we focused on the persistence dynamics of an
established invader, the European gypsy moth (Lymantria dispar) in the United States. The spread of gypsy
moth is largely thought to be the result of the growth
and coalescence of isolated colonies in a transition zone
ahead of the generally infested area. One important
question is thus the ability of these isolated colonies to
persist when subject to Allee effects and inimical stochastic events. We analyzed the US gypsy moth survey
data and identified isolated colonies of gypsy moth using
the local indicator of spatial autocorrelation. We then
determined region-specific probabilities of colony persistence given the population abundance in the previous
year and its relationship to a suite of ecological factors.
We observed that colonies in Wisconsin, US, were significantly more likely to persist in the following year
than in other geographic regions of the transition zone,
and in all regions, the abundance of preferred host tree
species and land use category did not appear to influence
persistence. We propose that differences in region-specific rates of persistence may be attributed to Allee
effects that are differentially expressed in space, and that
Communicated by John Reeve
S. L. Whitmire
Department of Biology, West Virginia University,
P. O. Box 6057, Morgantown, WV 26506, USA
E-mail: [email protected]
P. C. Tobin (&)
USDA Forest Service, Northeastern Research Station,
180 Canfield Street, Morgantown, WV 26505-3180, USA
E-mail: [email protected]
Tel.: +1-304-2851514
Fax: +1-304-2851505
the inclusion of geographically varying Allee effects into
colony-invasion models may provide an improved paradigm for addressing the establishment and spread of
gypsy moth and other invasive exotic species.
Keywords Persistence Æ Biological invasions Æ Allee
effects Æ Invasive species Æ Local indicator of spatial
autocorrelation
Introduction
Persistence is the ability of a population to sustain itself
in the environment subsequent to its arrival. Because of
the role it plays in the continuity of populations, persistence is of central importance to studies in epidemiology, conservation biology, and invasion ecology (Shea
1998; Earn et al. 1998). In epidemiological and conservation biological investigations, much interest has been
placed on the spatiotemporal dynamics of populations,
in which persistence may or may not be enhanced
depending on its relationship to population synchrony
or asynchrony (Heino et al. 1997; Rohani et al. 1999).
Due to the threat that invasive species pose to biological
diversity, considerable research efforts and resources are
devoted to documenting, predicting, and managing their
effects (Pimentel et al. 2000). As such, the invasion
dynamics of alien species in new habitats motivates key
questions regarding persistence.
Biological invasions occur in three stages: arrival,
establishment, and spread (Hengeveld 1988). Once an
alien species is established in a new region, the prevention of local spread is often difficult, or logistically
unrealistic (Krushelnycky et al. 2004). Population
spread of an established invader can occur either by
short-distance dispersal leading to continuous expansion
of an established population, or by long-distance dispersal, which could be aided by human transport of
propagules (Shigesada and Kawasaki 1997; Shigesada
et al. 1995). Long-distance spread can result in the formation of isolated colonies ahead of areas that are
231
generally infested. If these colonies are successful in
persisting, then they can increase in abundance, coalesce, and eventually become incorporated into the range
of the established population. In this manner, the overall
spread of an invading organism can be accelerated
(Sharov and Liebhold 1998b).
Low density, isolated colonies are often subject to
Allee effects and stochasticity, both of which can inhibit
the colonies’ ability to persist (Liebhold and Bascompte
2003; Hastings 1996) and thus considered important
factors when predicting the likelihood of the extinction
of small colonies (Leung et al. 2004; Liebhold and
Bascompte 2003). Allee effects collectively describe a
decrease in fitness with a decrease in population abundance, and causes include the inability to locate mates,
inbreeding depression, and failure to satiate predators
(Allee et al. 1949; Stephens et al. 1999). Environmental
stochasticity results in variable reproductive and mortality rates, causing population growth rates to vary.
Laboratory experiments in which environmental variation was manipulated resulted in an increase in colony
extinction and decrease of new colony establishment
with increasing levels of stochastic variation (Drake and
Lodge 2004). Stochastic events can have a strong influence on colony persistence, particularly for low density,
isolated colonies.
Low-density colonies could persist as a result of
metapopulation dynamics (Hanski 1999), in which
metapopulations sustain themselves due to a balance
between random extinction and recolonization of patches through dispersal (Levins 1969). Persistence of
metapopulations has been shown to be influenced by
population size within the patch (Berger 1990; Moilanen
and Hanski 1998), age of the patch (Hastings 2003),
spatial and landscape structure (With 2004; Hanski
1994; Moilanen and Hanski 1998), and environmental
and demographic stochasticity (Fagan 1999). Moreover,
populations governed by nonlinear dynamics, in which
local population noise is amplified, may experience decrease in extinction rates due to increase in population
asynchrony (Allen et al. 1993). There are numerous
examples of invasions of exotic species where a small
isolated colony persisted for some time prior to population expansion and consequent explosion (e.g., Shigesada and Kawasaki 1997, and examples therein). We
sought to investigate the ability of isolated colonies of an
established invasive species to persist ahead of a moving
population front. We used the invasion of the European
strain of the gypsy moth, Lymantria dispar (L.) (Lepidoptera: Lymantriidae), in the United States as a model
system for examining this type of persistence dynamics.
The gypsy moth was introduced into North America
in the 1860s near Boston, MA, USA. Its current range
includes southern Canada, west to Wisconsin, and south
to North Carolina. The gypsy moth is a polyphagous
herbivore with over 300 species of deciduous and
coniferous hosts (Elkinton and Liebhold 1990). Since
1924, over 34.6 million hectares of US forests have been
partially or completely defoliated by gypsy moth (Gypsy
Moth Digest 2004). The ecological and economic costs,
both indirect and direct, have been well documented
(e.g., Doane and McManus 1981; Thurber et al. 1994;
Leuschner et al. 1996; Sample et al. 1996; Redman and
Scriber 2000; Mayo et al. 2003).
Overwintering eggs hatch in the spring, and larvae
feed on both young and fully developed foliage. Pupation occurs mainly on the trunks of trees, and adults
emerge from mid to late summer. The gypsy moth is
univoltine, and in the European strain of gypsy moth,
females do not fly and usually oviposit within 1–2 m
from the site of adult emergence (Odell and Mastro
1980). Forms of dispersal include ballooning early instars, and adult male flight. Gypsy moth population
dynamics are complex (Campbell 1967, 1973; Doane and
McManus 1981; Elkinton and Liebhold 1990; Williams
and Liebhold 1995). Dwyer et al. (2004) recently suggested a host-pathogen-predator model, coupled with
environmental stochasticity, to explain the gypsy moth
outbreaks. In outbreak populations, viral epizootics can
regulate gypsy moth populations though they impose
little influence in low-density populations (Doane 1970;
Elkinton and Liebhold 1990). Instead, low-density
populations are most likely to be influenced by small
mammals (Bess et al. 1947; Campbell et al. 1977; Elkinton et al. 1989, 2004; Elkinton and Liebhold 1990),
such as the white-footed mouse, Peromyscus leucopus,
whose populations in turn can be influenced by mast
dynamics (Elkinton et al. 1996; Jones et al. 1998).
Regions of the US can be characterized as uninfested
by the gypsy moth, generally infested, or within some
transition zone between the two. In the transition zone,
the gypsy moth invades previously uninfested habitat,
initiates and establishes new colonies, which in turn can
serve as a source for new infestations. Gypsy moth
movement within this transition zone can vary from
year-to-year and from area to area. For example, gypsy
moth spread in the Appalachian states of West Virginia
and Virginia has averaged roughly 2 km/year
(SD=20.9 km) since 1996, while the corresponding rate
of spread in Wisconsin was 20 km/year (SD=34.4 km),
even though the management tactics in both regions
were uniformly applied under the USDA Forest Service
Slow-the-Spread program (Sharov et al. 2002; Tobin
et al. 2004). Past analyses on gypsy moth spread and the
factors that influence it have occasionally led to conflicting conclusions. For example, Liebhold et al. (1992)
reported that in colder climates, such as those areas
where minimum January temperatures were <7C,
historical rates of spread in the Northeastern US were
roughly three times slower than under the inverse condition. In contrast, Sharov et al. (1999) reported that in
Michigan, rates of spread were actually higher in colder
climates and instead linked the higher rates of spread to
the increases in forest susceptibility.
In the transition zone, the gypsy moth populations do
not necessarily spread continuously along the population front. Rather, individual colonies become established beyond the expanding front. Adult dispersal is
232
limited, and ballooning instars are usually deposited
within only a few hundred meters from the source
(Mason and McManus 1981). Thus, the movement of
gypsy moth beyond the infested zone is largely thought
to be the result of the inadvertent transportation of life
stages by humans (Schwalbe 1981; Mason and McManus 1981; Liebhold et al. 1992). Indeed, many new
infestations have been positively associated with household moves from infested to uninfested zones (McFadden and McManus 1991). These patterns suggest that
within the transition zone, colonies coalesce and contribute to the range expansion of gypsy moth (Sharov
and Liebhold 1998b).
In support of this argument, Liebhold et al. (1992)
modeled the gypsy moth range expansion and contended
that if windborne instar movement was the only form of
dispersal, then range expansion should occur at a rate of
roughly 3 km/year. However, they estimated that the
average spread rate was 21 km/year from 1965 to
1990, and they suggested that this greater rate of spread
was due, in part, to the formation, growth, and coalescence of isolated populations ahead of the population
front. However, it is unknown whether all isolated gypsy
moth colonies, particularly low-density ones, persist
from year-to-year within this transition zone, and if
there is a regional effect to these invasion dynamics. We
examined these questions by analyzing the persistence of
isolated colonies from year-to-year over the spatial extent of the transition zone from 1996 to 2003 given the
initial density of the colony and its relationship to a suite
of ecological factors.
Materials and methods
Sampling regime
Gypsy moth populations within the transition zone are
monitored through the USDA Forest Service Slow-theSpread Program (Sharov et al. 2002; Tobin et al. 2004).
One aspect of this program is to delineate population
boundaries, as well as boundaries that delimit the transition zone (Sharov et al. 1995). The boundaries of the
transition zone, which is roughly 100 km in width, are
then delineated such that the boundary closest to the
advancing population front is approximately 30–50 km
from the 10-moth population boundary (Tobin et al.
2004). In the transition zone, pheromone-baited traps
are placed at an intertrap distance of 2 km, which has
been shown to be sufficient for detecting isolated colonies (Sharov and Liebhold 1998a). The traps are placed
using handheld GPS units within a 500 m radius from
target coordinates. From year-to-year, the number of
traps placed in the transition zone varies, but generally
exceeds 50,000. We used the gypsy moth pheromone
trap data from 1996 to 2003 to test the spatial pattern of
isolated gypsy moth colony persistence. During these
years, traps within the transition zone were placed in
Wisconsin, Illinois, Indiana, Ohio, West Virginia, Virginia and North Carolina.
Identification of isolated colonies
We used Local Indicator of Spatial Autocorrelation
methods to estimate the local Moran statistic (Anselin
1995; Getis and Ord 1996; Boots 2002) in R (R Development Core Team 2002) using the spdep package
(Bivand 2004) to objectively identify spatially unique
colonies of the gypsy moth. For each year, the transition
zone was divided into 150·150 km2 blocks. In each
block, we estimated the local Moran statistic at each
trap according to
local Moran ¼
N
X
ðzi ZÞ
wij ðzi ZÞ
varðZÞ j¼1
ð1Þ
where zi is the number of moths recorded by the trap, zj
is the number of moths recorded by a trap within a local
neighborhood around zi, and Z and var (Z) refer to the
block mean and variance, respectively. Prior to estimation, the moth counts were transformed using
log10(z+1). The weight function, wij, is binary and equal
to 1 when a neighboring trap (i.e., zj) is located within
some a priori distance of zi. Around each zi, we used a
radius of 3.1 km to capture these neighboring traps.
Negative local Moran values represent traps that are
locally spatially different from the value of their neighbors (i.e., ‘‘high–low’’ or ‘‘low–high’’ associations, cf.
Anselin 1995). We considered traps across all years for
which the local Moran statistic was significantly negative
(Getis and Ord 1996). Because negative local Moran
values can result from either ‘‘high–low’’ or ‘‘low–high’’
associations, we further partitioned the data by selecting
the former association. Other traps were eliminated
from further analysis if they were in an area treated for
gypsy moth or if they fell outside the transition zone.
Lastly, we considered only traps for which the initial
moth abundance was ‡1 and £ 20. This final subset of
low-density, isolated colonies was then used in subsequent analyses.
Modeling persistence
For our response variable, we scored isolated colonies as
extinct if in year t+1 the trap catch for the colony was 0,
and persisting otherwise. Because exact trap locations
can move from year-to-year, we matched the trap location in year t to its closest location in year t+1. [The
mean distance between traps in successive years was
within the 500 m buffer within which the traps are
placed, and the distances did not differ significantly
among regions in an analysis of variance (F=1.6, df=2,
873, P=0.2)]. We used logistic regression (PROC LOGISTIC, SAS Institute 1999) to test the main effects of
the (1) abundance of the colony in year t; (2) the region
233
in which the colony was located; (3) abundance of moths
recorded from neighboring traps within 3.1 km of the
colony in year t; (4) the distance of the colony to the
generally infested area; (5) the basal area of the gypsy
moth preferred host species associated with the colony;
(6) USGS land cover category associated with the colony; and (7) interactions between main effects. We used
Hosmer and Lemeshow (2001) model selection methods
to determine the appropriate logistic regression model.
The significance of main and interaction effects were
based on the Wald v2 for Type 3 analysis. When
appropriate, odds ratios and associated confidence
intervals were estimated based on the Wald v2.
Colonies were placed into region categories as follows: Wisconsin (WI); the Midwest United States (MW):
traps in Illinois (IL), Indiana (IN), and Ohio (OH); the
Appalachian region of the United States (AP): traps in
West Virginia (WV), Virginia (VA), and North Carolina
(NC). The basal area of the gypsy moth preferred host
species was adjusted for forest density (Morin et al.
2005). Land cover data were obtained from the National
Land Cover Data (USGS 2003) at a 30 m resolution. All
forest types were collapsed into one forest category, as
were all agricultural areas.
Additional analyses
To examine neighborhood effects on persistence, we
modified our response variable to include moth captures
from traps within a prespecified area around the isolated
colony in year t+1. In this case, if the sum of moths
caught at the isolated colony and in the traps in the
neighborhood around the isolated colony was 0 in year
t+1, then we considered the colony extinct; otherwise,
the colony was considered to be persisting. We used two
radii, 3.1 and 6.2 km, around the isolated colony to
define these neighborhoods. Lastly, we examined the
effect of elevation on isolated colony persistence in AP,
where there are reasonable elevation differences. Elevation data were obtained from the North America digital
elevation model, GTOP030. GTOPO30 is a digital elevation model (DEM) with a horizontal grid spacing of
30 arc s (1 km) (USGS 2004).
Results
Across all years, the number of isolated colonies detected by the local Moran statistic in WI, MW, and AP
were 321, 231, and 324, respectively. The local Moran
method was successful in identifying the isolated colonies: across all regions; for example, over 50% of the
colonies were surrounded by neighboring traps (i.e.,
those within 3.1 km radius of the colony) in which no
moths were caught, and over 90% were surrounded by
neighboring traps in which total moth counts were £ 2.
We observed significant effects of the (1) abundance of
the colony in year t (v2=6.9, df=1, P<0.01, Figs. 1, 2)
region in which the colony was located (v2=39.2, df=2,
P<0.01, Fig. 1); (3) abundance of moths recorded from
neighboring traps within 3.1 km of the colony in year t
(v2=4.1, df=1, P=0.04); and (4) the distance of the
colony to the generally infested area (v2=24.4, df=1,
P<0.01, Fig. 2). No other main effects or interactions
between main effects were significant (Hosmer and
Lemeshow v2=4.3, df=8, P=0.83). The isolated colonies in WI were 1.7 (95% CI, 1.2–2.3) and 2.5 (1.7–3.6)
times more likely to persist than colonies in AP and MW,
respectively, while the AP and MW were not significantly
different (Fig. 1). Collapsing over initial abundance, the
mean (variance) overall probability of persisting from
year-to-year in WI, AP, and MW was 0.492 (0.250),
0.350 (0.228), and 0.268 (0.196), respectively.
Isolated colonies were more likely to persist when
closer to the generally infested area in all regions
(Fig. 2). There was no interaction effect between the
region and distance from the generally infested zone
(v2=1.4, df=2, P=0.50), suggesting slope homogeneity
among regions. However, there was a significant region
effect (v2=53.5, df=2, P<0.01), such that at approximately 140 km from the generally infested area, the
colonies in WI persisted roughly 50% of the time, while
in AP and MW the corresponding rates were 30 and
20%, respectively.
The basal area of preferred host species and land use
category did not explain differences in persistence rates
among regions (Fig. 3). It was evident that even in areas
of high host tree abundance, persistence of isolated gypsy
moth colonies was not necessarily enhanced. Also, most
of the colonies in all regions were located in areas classified as either forest or agricultural lands, with only a few
colonies located in wetlands or urban areas (Table 1). In
WI, isolated gypsy moth colonies still persisted roughly
half of the time regardless of the land use category in
which the colony was located (Table 1). In AP and MW,
the colonies persisted approximately 35 and 20% of the
time, respectively, and likewise regardless of the land use
category (Table 1). However, it is important to note that
there were differences in scale between the gypsy moth
trap catch data, and in the scale used to interpolate the
basal area of the gypsy moth preferred host species
(Morin et al. 2005) and NLCD data (USGS 2003), which
could render associations difficult to ascertain.
We observed that when the moths measured by
neighboring traps in year t+1 are included in our criterion of colony persistence, persistence in all regions
increased relative to their exclusion (Fig. 4). However,
the regions remained significantly different from each
other in the same fashion as previously observed. The
results were similar when using either a 3.1 or 6.2 km
radius. For a 3.1 km radius, isolated colonies in WI were
2.2 (95% CI, 1.5–3.3) and 3.6 (2.4–5.3) times more likely
to persist than those in AP and MW, respectively, while
colonies in AP were 1.6 (1.1–2.3) times more likely to
persist than those in MW.
The range of elevation in AP was 0–1,132 m above
sea level. In this region, 41% of the colonies at
234
Fig. 1 Predicted probabilities of the gypsy moth (Lymantria dispar)
colony persistence based on the abundance of the colony from the
preceding year in a All regions, b Appalachia (AP), c Midwest
(MW), and d Wisconsin (WI). The histograms represent the
observed frequency of colonies that became extinct (black bars)
or persisted (gray bars) (cf. Smart et al. 2004). The model
predictions (solid line) and 95% confidence intervals (dashed lines)
are also shown for each region
Table 1 Percentage (and total number) of isolated gypsy moth colonies persisting (P) or not (E) by land use category and region
Region
Land use
Category
Forest
Agriculture
Wetland
Urban
AP
P
33.3
39.3
50.0
57.1
(80)
(24)
(5)
(4)
E
66.7
60.7
50.0
42.9
(160)
(37)
(5)
(3)
MW
P
23.4 (11)
28.5 (47)
0.0 (0)
21.1 (4)
E
76.6
71.5
0.0
79.0
(36)
(118)
(0)
(15)
WI
P
46.7
49.6
53.9
71.4
(70)
(69)
(14)
(5)
E
53.3
50.4
46.2
28.6
(80)
(70)
(12)
(2)
The consistency in persistence/extinction between forest and agricultural areas for each region. Data from wetland and urban areas are
shown but were not analyzed due to insufficient sample sizes
elevations <200 m persisted while only 25% at elevations 200–400 m persisted, and these rates were significantly different (G2=6.2, df=1, P=0.01). Furthermore,
most (72%) of the variation in the elevation effect was
explained by the differences between colonies at elevations <200 m and those at 200–400 m. Sharov et al.
(1997) reported that in uninfested areas, higher colonization rates were higher at lower elevation, possibly due
to increased human movement, and this study supports
the notion of enhanced persistence at lower elevations.
Discussion
The difference in region-specific rates of the gypsy moth
colony persistence within the transition zone has
important consequences in the dynamics of gypsy moth
range expansion, and implications for the dynamics of
other invading species. Low-density, isolated populations are generally subject to Allee effects, which can
affect their ability to establish, spread, and persist (Lewis
and Kareiva 1993; Hastings 1996; Keitt et al. 2001;
Fagan et al. 2002; Liebhold and Bascompte 2003; Drake
2004; Drake and Lodge 2005). Indeed, the gypsy moth
colonies in all regions that we explored generally had
lower probabilities of persisting at lower densities
(Fig. 1). Liebhold and Bascompte (2003) illustrated the
importance of Allee effects in the isolated colonies of
gypsy moth in an uninfested area of the United States.
In this case, life stages were unintentionally introduced,
and they contended that the failure of these colonies to
persist and become established was due to a failure of
finding mates. In our study, we propose that not only are
Allee effects important in the gypsy moth persistence,
235
Fig. 2 Region-specific predicted probabilities (±95% confidence
intervals) of persistence in year t+1 of isolated colonies, regardless
of their initial abundance in year t, over the distance between the
colony and the gypsy moth generally infested area for a Appalachia
(AP), b Midwest (MW), and c Wisconsin (WI)
but that the strength of the Allee effect appears to vary
over geographic regions. This work shows how it thus
may be important to consider region-specific Allee effects when addressing biological invasions.
In many invading species, dispersal from initial parent populations subsequent to invasion may be limited.
The gradual range expansion generally follows a colonyinvasion model in which spread can be accelerated by
the coalescence of isolated colonies, or propagules, that
establish away from the initial population (Shigesada
and Kawasaki 1997). The inadvertent movement of life
stages by humans is thought to be a key mechanism by
which many species, including the gypsy moth (Schwalbe 1981; Mason and McManus 1981; Liebhold et al.
1992), traverse great distances. Moreover, recent work
on the horse chestnut leafminer underscores the importance of the relationship between human population
density and long-distance movement by an invading
insect (Gilbert et al. 2004).
Fig. 3 Histogram of the frequency of isolated gypsy moth colonies
that became extinct (black bars) or persisted (gray bars) over the
quantity (basal area) of preferred gypsy moth host trees in which
the colony was located for a Appalachia (AP), b Midwest (MW),
and c Wisconsin (WI)
The inclusion of spatially explicit Allee effects into the
colony-invasion models seems to be applicable for the
gypsy moth, and may also be applicable for other
invading species. One result of such a paradigm could be
the development of irregular population spread boundaries from an initial infestation point, in which propagules in space are subject to varying Allee effects.
Biological invasions with such corrugated boundaries
have been hypothesized to increase their range expansion more rapidly than those with planer boundaries
(Lewis and Kareiva 1993). The gypsy moth provides an
example of this phenomenon. In WI, the spread of
invading gypsy moth does often lead to the appearance
of ‘‘bulges’’ in population boundaries, whereas in other
regions, the spread follows a more planar boundary
(Decision-Support System for the Slow-the-Spread
Project 2005). In this study, we have observed regional
differences in colony persistence, perhaps due to geographically varying Allee effects. If such Allee effects are
236
Fig. 4 Effect of the size of the local neighborhood to region-specific
predicted probabilities of persistence of the colony and its
neighborhood (within 3.1 or 6.2 km) given their abundance from
the preceding year. In the AP and MW regions, this relationship
did not differ significantly when using the 6.2 km neighborhood
occurring in the spread dynamics of the gypsy moth,
then this may provide a plausible explanation for the
more rapid rate of spread in WI relative to other regions
of the United States (Decision-Support System for the
Slow-the-Spread Project 2005). However, in the case of
gypsy moth, it is still not clear which process, persistence
or spread, is driving the other.
In this study, we demonstrated the potential for Allee
effects to be differentially expressed in space. These
findings have implications in the development of population dynamics and spread models in both established
invaders and newly arrived invaders. There are also
important implications in the overall management of
invasive species, such as the importance of management
protocols to be specific to the locality of infestation, as
well as underscoring the uncertainty associated with
management over regional scales.
Acknowledgments We thank Sandy Liebhold, Randy Morin, and
Gino Luzader (USDA Forest Service), Desta Fekedulegn and
Jonathan Cumming (West Virginia University Statistics and Biology, respectively), and Andy Roberts and Matt Learn (Virginia
Tech University Entomology) for their assistance. We also thank
Sandy Liebhold and Mary Ann Fajvan for their critical comments
in the preparation of this manuscript. This research was supported
by USDA Forest Service Research Cooperative grant no. 04-CA11242343-063.
References
Allee WC, Emerson AE, Park O, Park T, Schmidt KP (1949)
Principles of animal ecology. WB Saunders, Philadelphia
Allen JC, Schaffer WM, Rosko D (1993) Chaos reduces species
extinction by amplifying local population noise. Nature
364:229–232
Anselin L (1995) Local Indicators of Spatial Association - Lisa.
Geogr Anal 27:93–115
Berger J (1990) Persistence of different-sized populations: An
empirical assessment of rapid extinctions in bighorn sheep.
Conserv Biol 4:91–98
Bess HA, Spurr SH, Littlefield EW (1947) Forest site conditions
and the gypsy moth. Harv For Bull 22:1–26
Bivand R (2004) Spatial dependence: weighting schemes, statistics,
and models. http://cran.r-project.org/doc/packages/spdep.pdf.
(cited 15 July 2004)
Boots B (2002) Local measures of spatial association. Ecoscience
9:168–176
Campbell RW (1967) The analysis of numerical change in gypsy
moth populations. For Sci Monogr 15:1–33
Campbell RW (1973) Numerical behavior of a gypsy moth population system. For Sci 19:162–167
Campbell RW, Sloan RJ, Biazak CE (1977) Sources of mortality
among late instar gypsy moth (Lepidoptera : Lymantriidae)
larvae in sparse populations. Environ Entomol 6:865–871
Decision-Support System for the Slow-the-Spread Project (2005)
http://da.ento.vt.edu/. Cited 6 June 2005
Doane CC (1970) Primary pathogens and their role in development
of an epizootic in gypsy moth. J Invertebr Pathol 15:21–27
Doane CC, McManus ME (eds) (1981) The gypsy moth: research
toward integrated pest management. USDA Tech Bull 1584,
Washington
Drake JM (2004) Allee effects and the risk of biological invasion.
Risk Anal 24:795–802
Drake JM, Lodge DM (2004) Effects of environmental variation on
extinction and establishment. Ecol Lett 7:26–30
Drake JM, Lodge DM (2005) Allee effects, propagule pressure and
the probability of establishment: risk analysis for biological
invasions. Biol Invasions (in press)
Dwyer G, Dushoff J, Yee SH (2004) The combined effects of
pathogens and predators on insect outbreaks. Nature 430:341–
345
Earn DJD, Rohani P, Grenfell BT (1998) Persistence, chaos and
synchrony in ecology and epidemiology. Proc R Soc Lond B
Biol Sci 265:7–10
Elkinton JS, Gould JR, Liebhold AM, Smith HR, Wallner WE
(1989) Are gypsy moth populations regulated at low density?
In: Wallner WE, McManus KA (eds) Lymantriidae: a comparison of features of new and old world tussock moths. USDA
For Serv GTR-NE-123, Washington, pp 233–249
Elkinton JS, Liebhold AM (1990) Population dynamics of gypsy
moth in North America. Annu Rev Entomol 35:571–596
Elkinton JS, Healy WM, Buonaccorsi JP, Boettner GH, Hazzard
A, Liebhold AM, Smith HR (1996) Interactions among gypsy
moths, white-footed mice, and acorns. Ecology 77:2332–2342
Elkinton JS, Liebhold AM, Muzika RM (2004) Effects of alternative prey on predation by small mammals on gypsy moth
pupae. Popul Ecol 46:171–178
Fagan WF (1999) Weak influences of initial conditions on metapopulation persistence times. Ecol Appl 9:1430–1438
Fagan WF, Lewis MA, Neubert MG, van den Driessche P (2002)
Invasion theory and biological control. Ecol Lett 5:147–157
Getis A, Ord JK (1996) Local spatial statistics: an overview. In:
Longley P, Batty M (eds) Spatial analysis: modelling in a GIS
environment. Wiley, New York, pp 261–277
Gilbert M, Gregoire J-C, Freise JF, Heitland W (2004) Long-distance dispersal and human population density allow the prediction of invasive patterns in the horse chestnut leafminer
Cameraria ohridella. J Anim Ecol 73:459–468
Gypsy Moth Digest (2004) USDA Forest Service, State and Private
Forestry, Morgantown, WV http://na.fs.fed.us/wv/gmdigest/
index.html (cited 15 July 2004)
Hanski I (1994) A Practical Model of Metapopulation Dynamics. J
Anim Ecol 63:151–162
Hanski I (1999) Habitat connectivity, habitat continuity, and
metapopulations in dynamic landscapes. Oikos 87:209–219
Hastings A (1996) Models of spatial spread: a synthesis. Biol
Conserv 78:143–148
Hastings A (2003) Metapopulation persistence with age-dependent
disturbance or succession. Science 301:1525–1526
Heino M, Kaitala V, Ranta E, Lindström J (1997) Synchronous
dynamics and rates of extinction in spatially structured populations. P Roy Soc Lond B Biol 264:481–486
Hengeveld R (1988) Mechanisms of Biological Invasions. J Biogeogr 15:819–828
237
Hosmer DW, Lemeshow S (2001) Applied logistic regression, 2nd
edn. Wiley, New York
Jones CG, Ostfeld RS, Richard MP, Schauber EM, Wolff JO
(1998) Chain reactions linking acorns to gypsy moth outbreaks
and Lyme disease risk. Science 279:1023–1026
Keitt TH, Lewis MA, Holt RD (2001) Allee effects, invasion pinning, and species’ borders. Am Nat 157: 203–216
Krushelnycky PD, Loope LL, Joe SM (2004) Limiting spread of a
unicolonial invasive insect and characterization of seasonal
patterns of range expansion. Biol Invasions 6:47–57
Leung B, Drake JM, Lodge DM (2004) Predicting invasions:
Propagule pressure and the gravity of allee effects. Ecology
85:1651–1660
Levins R (1969) Some demographic and genetic consequences of
environmental heterogeneity for biological control. Bull Entomol Soc Am 15: 237–240
Liebhold A, Bascompte J (2003) The Allee effect, stochastic
dynamics and the eradication of alien species. Ecol Lett 6:133–
140
Liebhold AM, Halverson JA, Elmes GA (1992) Gypsy-moth
invasion in North-America - a quantitative-analysis. J Biogeogr
19:513–520
Leuschner WA, Young JA, Walden SA, Ravlin FW (1996) Potential benefits of slowing the gypsy moth’s spread. South J
Appl For 20: 65–73
Lewis MA, Kareiva P (1993) Allee dynamics and the spread of
invading organisms. Theoret Pop Biol 43:141–158
Mason CJ, McManus ML (1981) Larval dispersal of the gypsy
moth. In: Doane CC, McManus ML (eds) The gypsy moth:
Research toward integrated pest management. USDA Tech
Bull 1584, Washington, pp 161–202
Mayo JH, Straka TJ, Leonard DS (2003) The cost of slowing the
spread of the gypsy moth (Lepidoptera: Lymantriidae). J Econ
Entomol 96:1448–1454
McFadden MW, McManus ME (1991) An Insect out of control?
The potential for spread and establishment of the gypsy moth in
new forest areas in the United States. In: Baranchikov YN,
Mattson WJ, Hain FP, Payne TL (eds) Forest insect guilds:
patterns of interaction with host trees, GTR-NE-153. USDA
Forest Service, Washington, pp 172–186
Moilanen A, Hanski I (1998) Metapopulation dynamics: effects of
habitat quality and landscape structure. Ecology 79:2503–2515
Morin RS, Liebhold AM, Luzader ER, Lister AJ, Gottschalk KW,
Twardus DB (2005) Mapping host-species abundance of three
major exotic forest pests USDA research Paper NE-726, USDA
Forest Service, Newtown Square, Pennsylvania
Odell TM, Mastro VC (1980) Crepuscular activity of gypsy moth
adults (Lymantria dispar). Environ Entomol 9:613–617
Pimentel D, Lach L, Zuniga R, Morrison D (2000) Environmental
and economic costs of nonindigenous species in the United
States. Bioscience 50:53–65
R Development Core Team (2004) http://www.r-project.org (cited
9 July 2004)
Redman AM, Scriber JM (2000) Competition between the gypsy
moth, Lymantria dispar, and the northern tiger swallowtail,
Papilio canadensis: interactions mediated by host plant chemistry, pathogens, and parasitoids. Oecologia 125:218–228
Rohani P, Earn DJD, Grenfell BT (1999) Opposite patterns of
synchrony in sympatric disease metapopulations. Science
286:968–971
Sample BE, Butler L, Zivkovich C, Whitmore RC, Reardon R
(1996) Effects of Bacillus thuringiensis Berliner var. kurstaki
and defoliation by the gypsy moth, Lymantria dispar (L.)
(Lepidoptera: Lymantriidae) on native arthropods in West
Virginia. Can Entomol 128:573–592
SAS Institute (1999) SAS/STAT User’s guide, Version 8. Cary,
NC
Schwalbe CP (1981) Disparlure-baited traps for survey and detection. In: Doane CC, McManus ML (Eds) The gypsy moth research toward integrated pest management. USDA Tech Bull
1584, Washington, pp. 542–548
Sharov AA, Liebhold AM (1998a) Quantitative analysis of gypsy
moth spread in the Central Appalachians. In: Baumgartner,
et al (eds) Population and Community Ecology for Insect
Management and Conservation. Balkema, Rotterdam, pp 99–
110
Sharov AA, Liebhold AM (1998b) Model of slowing the spread of
gypsy moth (Lepidoptera : Lymantriidae) with a barrier zone.
Ecol Appl 8:1170–1179
Sharov AA, Pijanowski BC, Liebhold AM, Gage SH (1999) What
affects the rate of gypsy moth (Lepidoptera: Lymantriidae)
spread: winter temperature or forest susceptibility?. Agric For
Entomol 1:37–45
Sharov AA, Leonard D, Liebhold AM, Roberts EA, Dickerson W
(2002) ‘‘Slow the Spread’’: a national program to contain the
gypsy moth. J Forest 100:30–35
Sharov AA, Liebhold AM, Ravlin FW (1995) Prediction of gypsymoth (Lepidoptera, Lymantriidae) mating success from pheromone trap counts. Environ Entomol 24:1239–1244
Sharov AA, Liebhold AM, Roberts EA (1997) Correlation of
counts of gypsy moths (Lepidoptera: Lymantriidae) in pheromone traps with landscape characteristics. Forest Sci 43:483–
490
Shea K (1998) Management of populations in conservation, harvesting and control. Trends Ecol Evol 13:371–375
Shigesada N, Kawasaki K (1997) Biological invasions: theory and
practice. Oxford University Press, Oxford
Shigesada N, Kawasaki K, Takeda Y (1995) Modeling stratified
diffusion in biological invasions. Am Nat 146:229–251
Smart J, Sutherland WJ, Watkinson AR, Gill JA (2004) A new
means of presenting the results of logistic regression. Bull Ecol
Soc Am 85: 100–102
Stephens PA, Sutherland WJ, Feckleton RP (1999) What is the
Allee effect? Oikos 87: 185–190
Thurber DK, McClain WR, Whitmore RC (1994) Indirect effects
of gypsy-moth defoliation on nest predation. J Wildl Manage
58:493–500
Tobin PC, Sharov AA, Leonard DS, Roberts EA, Liebhold AM
(2004) Management of the gypsy moth through a decision
algorithm under the Slow-the-Spread project. Am Entomol
50:200–209
USGS (2003) National Land Cover Data. http://landcover.usgs.gov/natllandcover.asp. (cited 9 July 2004)
USGS (2004) North America digital elevation model. http://edcdaac.usgs.gov/gtopo30/Readme.html (cited 20 June 2004)
Williams DW, Liebhold AM (1995) Influence of weather oil the
synchrony of gypsy-moth (Lepidoptera, Lymantriidae) outbreaks in New-England. Environ Entomol 24:987–995
With KA (2004) Assessing the risk of invasive spread in fragmented
landscapes. Risk Analysis 24:803–815